Investing in AI & Big Data 2021

Published on 03/20/21 | Saurav Sen | 2,139 Words

The BuyGist:

  • AI & Big Data is our biggest thematic bet – accounting for 30-40% of our portfolio at any given time.
  • In this article, we take a landscape view of the theme and zoom in to a list of investment ideas.
  • The idea is to have a “ready-to-go” list of AI & Big Data investments IF there is a big market correction. This is always a big IF.
  • But we’d like to ready, nonetheless.

Our AI & Big Data Watch List

We maintain Watch Lists for many of our major investment themes, such as 5G, Media & Entertainment or Autonomous & Electro Mobility. AI & Big Data is our largest Watch List. It is also the largest thematic exposure in our portfolio.

Our Watch Lists are not algorithmic. There’s a fair amount of subjectivity built into them. All our Watch Lists contain companies that, we believe, have compelling stories. Either they are 1) Global Dominators or 2) have exponential growth trajectories. Some companies have both attributes. Some even have a third attribute: they show up on our other Watch Lists as well, meaning they have multiple tailwinds behind them.

Our Watch Lists have one purpose above all – they inform us if there’s anything reasonably priced in a particular theme. What is “reasonably priced”? We have a simple metric for reasonableness: we back-solve into what revenue growth we’d need to believe to buy the stock. This is an unusual way to screen investment ideas. But we think it’s the most logical, and it’s closest to our investment methodology. The name of the process is Expectations Investing – as coined by Michael Mauboussin. But we’ve adapted his teachings and combined it with our experience in Value Investing.

Here a snapshot of our full AI & Big Data Watch List:

We deliberately left out Social Media companies like Facebook from this list. We have a separate Media & Entertainment Watch List. But, yes, you could argue that they are efficient wielders of Big Data, using AI. In this article, we’ll focus on more boring but arguably more important companies.

The last column in the Watch List is the most important one – “Revenue Growth we need to Believe”. We’ve explained our methodology in detail here. But the gist is this: We back-solve into revenue growth rate based on some fixed assumptions:

  1. We never pay more than 20X Expected Free Cash Flow for a business.
  2. We assume current Gross Margins will remain the same in the future.
  3. We assume current Fixed Operating Costs will remain the same in the future.
  4. We always look for a 30% Margin of Safety – difference between current stock price and purchase price.

“Revenue growth we’d need to believe in order to buy the stock” is a cumulative number, not an annual growth rate. This is important. Now, you may ask, “cumulative for how long?”. We’d say, “an average of 5 years”. We expect some companies to achieve this cumulative growth in 2 years while some will achieve it in 7. We can’t be precise about it, but it really does vary from company to company.

These other assumptions are somewhat crude. But for a first cut, this methodology is the most logical and effective way to screen investment candidates. It relieves us from the futile task of making precise forecasts. It gives us a quick way to spend our time wisely.

To really make educated assumptions about the trajectory of gross margins or fixed operating costs, we would need to dig into the story – 1) a firm’s competitive advantage, 2) the durability of that competitive advantage and 3) the management team’s strategy and investment plan. This analysis is usually done on a short list of highly promising candidates. That’s what this article is about.

We’ll start by splitting the AI & Big Data Watch List into sub-themes and then narrow down our list using some other metrics.

Splitting up the Watch List

AI & Big Data does sound like a buzzword with a bit of a “castles in the sky” connotation. We obviously believe it’s not. But we understand the skepticism. So, maybe it’s easier to see the practicality if we split up the list into sub-themes. Here’s what we have:

A few points to note:

  1. Some companies straddle multiple sub-themes. We like that. Examples: Amazon, Google, Lumentum, Snowflake.
  2. Some companies straddle multiple themes outside AI & Big Data. We like that too. ex. Qualcomm(5G), Infosys and Wipro (India).
  3. And finally, some companies have less than 100% “revenue growth we need to believe”! This, we really like!

We can narrow our list down to a handful of companies that meet all these criteria. This would be the list:

  1. Google
  2. Lumentum
  3. Microsoft
  4. Box
  5. Salesforce
  6. Snowflake

We already have Lumentum and Microsoft in our portfolio. So, that leaves:

  1. Google
  2. Box
  3. Salesforce.
  4. Snowflake

This is a good start but let’s look at this from another angle.

The Multi-Theme Factor

We love it when we find companies that straddle two or more of our investment themes. And we love it even more when their stocks trade at comfortable prices. Straddling multiple themes means that there are multiple tailwinds that the company can harness. More specifically to us, they reduce our risk – straddling multiple tailwinds reduces the probability of permanent loss.

Multi-theme companies trading at comfortable prices are hard to find. But wild market-wide selloffs happen every now and then. It helps to be ready to pounce when others are fearful. That’s where big returns accrue. That’s why we have these Watch Lists. In our AI & Big Data Watch List, there are a few companies that straddle two or more themes. Here they are:

  1. Google – also Media & Entertainment. Its main business is advertising.
  2. Apple – also Retail and Media & Entertainment.
  3. Adobe – also Media & Entertainment.
  4. Salesforce – also Retail
  5. Lumentum – primarily 5G
  6. Qualcomm – primarily 5G
  7. Shopify – primarily Retail
  8. TSMC – also 5G, Industry 4.0 and AVs & EVs. The make chips for all end-markets.
  9. Twilio – also Industry 4.0. This is a potentially lucrative end-market in the next 10 years.
  10. Infosys – also India
  11. Wipro – also India
  12. Snowflake – potentially also Industry 4.0

We already have Apple, Lumentum and TSMC in our portfolio. So, we’re down to 9 companies so far. That’s a lot. So, let’s put them through another test to prioritize our time.

The Scalability Factor

Scalability refers to the ability to scale up profit (in our view that’s free cash flow) at a much faster rate than revenue growth. This would happen if two basic criteria are met:

  1. Obviously, it should be a growing business…AND…
  2. Most of its operating costs should be FIXED, not variable.

We can’t really ascertain whether a business will grow for the next 5-10 years unless we dig into the story of the company. We can’t simply extrapolate past growth rates into the future. Businesses change. Management teams change. The world changes. But we can, just maybe, extrapolate a firm’s Fixed Costs into the future. These costs should be, well, fixed, even if the business grows and production increases.

Variable costs increase with volume of production. We use Cost of Goods Sold (as reported) as our proxy for Variable Costs. Revenue minus Cost of Goods Sold is Gross Profit. Gross Profit as a percent of revenue is Gross Margin. We believe that it’s reasonable to assume, for now, that Gross Margins remain similar over time. In other words, Cost of Goods Sold increases proportionately with Revenue.

Let’s look at our shortlist of 9 companies we mentioned in the last section. We’ll focus on 2 main numerical factors:

  1. Revenue growth we’d need to believe…generally lower the better
  2. Scalability factor – defined as “fixed costs as a percent of total operating costs”…higher the better…

Here's the data:

So, to simplify this, let’s combine these metrics into one simple number where higher is better. Let’s divide Fixed Costs % by Revenue growth we’d need to believe. Let’s call it the Priority Factor. For our 9 companies so far, here are the stats:

This is our Priority List. It’s a good guide to how we should spend our time digging into companies. Over the next few weeks, we’ll try to answer the question: Should we believe these revenue growth rates. The real hard work is in trying to size up a company’s competitive advantage, and, more importantly, the durability of that competitive advantage for the next 5 to 10 years. Then, any revenue growth estimate can be taken seriously.

Next week, we’ll follow up with and investment thesis on one of these companies on the top of the list: Qualcomm, Adobe or Salesforce.

We use cookies on this site to ensure the best service possible.